Is there a difference in gamut between the regular driver and the plugin?
The consensus on the Wiki is that there is NO significant difference in gamut between prints made through the 8 bit operating system level driver and the 16 bit Photoshop plugin. This is based on the following evidence:
1) John Hollenberg created custom profiles for both the driver and the plugin on Canon Heavyweight Satin Photo Paper using the Bill Atkinson 1728 patch target and Profilemaker Pro 5.08. He followed the procedure for creating custom profiles detailed in this Wiki section. While there were slight differences in the measurements for some of the patches, 90% were within 0.5 delta E 94, and the worst 10% had a delta E 94 of 1.2. The patch with the biggest difference between the two targets still only showed a delta E 94 of 2.4. For reference, delta E of 1.0 is the smallest difference that can be reliably seen visually by the human observers. Of course, the resulting profiles showed virtually identical gamut when examined in Colorthink, with a very slight advantage to the profile made through the driver (yup, that is the driver that had a very slightly larger gamut 734,000 vs. 724,000 per Colorthink Pro calculation).
2) Some of those who initially thought that the plugin had a larger gamut turned out to be using slightly different methods to print the targets for plugin vs. driver, or were creating “8 bit profiles” for the driver using Printfix Pro, but “16 bit” files for the plugin. While the number of bits in the LUT (= Look Up Table) of the profile may not be that significant, the 8 bit profiles used a coarser grid (9 grid points vs 33 grid points) for the portion of the profile used for softproofing (or to display the gamut in Colorthink), thus giving a false impression as to the actual gamut of the profile. Prints made through both the driver and the plugin were not compared.
3) One poster who initially thought that the plugin had a higher gamut did more extensive testing and discovered that the number of passes (8 vs. 12 vs. 16) had a significant effect on the gamut. In some of the earlier comparisons, the targets for the driver may have been printed with only 8 passes rather than 16—hardly a fair comparison.
4) It seems likely that Color Management was inadvertantly left on in the driver for some of the people who saw a larger gamut for the plugin. If color management is on, you may be creating a profile for a printer color space already somewhat limited in size. Again, correct procedure is documented in the section on creating custom profiles.
Tentative Conclusion: The gamuts of the driver and plugin appear nearly identical when custom profiles are made for each under highest quality and equal conditions. When printing targets this means:
Using 16 bit targets for both driver and plugin
Printing targets using 600 PPI
Printing targets using 16 passes
Making “16 bit profiles” for both driver and plugin
Setting unidirectional or bidirectional printing the same for both methods
When making actual prints to compare, the same 16 bit file should be used, and the settings should be identical to those used when the profile is made.
Is there a difference in gamut between the regular driver and the plugin?
The consensus on the Wiki is that there is NO significant difference in gamut between prints made through the 8 bit operating system level driver and the 16 bit Photoshop plugin. This is based on the following evidence:
1) John Hollenberg created custom profiles for both the driver and the plugin on Canon Heavyweight Satin Photo Paper using the Bill Atkinson 1728 patch target and Profilemaker Pro 5.08. He followed the procedure for creating custom profiles detailed in this Wiki section. While there were slight differences in the measurements for some of the patches, 90% were within 0.5 delta E 94, and the worst 10% had a delta E 94 of 1.2. The patch with the biggest difference between the two targets still only showed a delta E 94 of 2.4. For reference, delta E of 1.0 is the smallest difference that can be reliably seen visually by the human observers. Of course, the resulting profiles showed virtually identical gamut when examined in Colorthink, with a very slight advantage to the profile made through the driver (yup, that is the driver that had a very slightly larger gamut 734,000 vs. 724,000 per Colorthink Pro calculation).
2) Some of those who initially thought that the plugin had a larger gamut turned out to be using slightly different methods to print the targets for plugin vs. driver, or were creating “8 bit profiles” for the driver using Printfix Pro, but “16 bit” files for the plugin. While the number of bits in the LUT (= Look Up Table) of the profile may not be that significant, the 8 bit profiles used a coarser grid (9 grid points vs 33 grid points) for the portion of the profile used for softproofing (or to display the gamut in Colorthink), thus giving a false impression as to the actual gamut of the profile. Prints made through both the driver and the plugin were not compared.
3) One poster who initially thought that the plugin had a higher gamut did more extensive testing and discovered that the number of passes (8 vs. 12 vs. 16) had a significant effect on the gamut. In some of the earlier comparisons, the targets for the driver may have been printed with only 8 passes rather than 16—hardly a fair comparison.
4) It seems likely that Color Management was inadvertantly left on in the driver for some of the people who saw a larger gamut for the plugin. If color management is on, you may be creating a profile for a printer color space already somewhat limited in size. Again, correct procedure is documented in the section on creating custom profiles.
Tentative Conclusion: The gamuts of the driver and plugin appear nearly identical when custom profiles are made for each under highest quality and equal conditions. When printing targets this means:
When making actual prints to compare, the same 16 bit file should be used, and the settings should be identical to those used when the profile is made.
See also the discussion in this forum thread.
Update 9/20/07: Scott Martin of Onsight confirmed that there is no difference in gamut between the plugin and driver.